Will Claude keep running if you close your laptop or lock it? (And how to run Claude Code 24/7)

What keeps running when you close the lid

A sleeping laptop runs nothing. Every program freezes, so a Claude Code session on the laptop pauses when the lid closes and picks up again when you wake it. A request to Anthropic that was in flight at that moment may fail when the laptop comes back. Anthropic's docs say the same about Remote Control: if your laptop sleeps or your network drops, the session reconnects once the machine is back online (Claude Code docs).

Lock, sleep or close the lid: what happens to Claude?

Locking puts a password in front of the screen. It doesn't put the computer to sleep, so its programs keep running. Sleep is what stops work. A sleeping computer doesn't run your programs until you wake it, and it drops into sleep by itself once its idle timer runs out.

Closing the lid is the fast way to sleep a laptop. Apple says closing the display puts a Mac laptop to sleep (Apple Support). Microsoft says a lot of Windows PCs, especially laptops, do the same when you close the lid (Microsoft Support). So for anything running on your own computer, one question decides it: will the computer sleep?

The Claude app is where people get caught out. Some of its work runs on Anthropic's servers and some runs on your computer, and the two behave differently. The sleep column is what Anthropic's help center and docs say, as of October 2026. The lock column is our reading of them, and the paragraph after the table explains why.

Anthropic moved new Cowork tasks on Pro and Max plans to its cloud. They keep going when you close your laptop, and their scheduled tasks run with no device online (Claude Help Center). Tasks you'd already started on your computer before that move stay there, and Anthropic says those may end if the app closes or the computer sleeps (Claude Help Center). A scheduled task that needs local files or apps only runs locally (Claude Help Center). On Team and Enterprise plans, Cowork in the cloud is still in beta.

Claude Code Desktop is stricter. Its local scheduled tasks only fire while the app is open and the computer is awake. If the computer sleeps through a run, it's skipped, and on wake the app starts one catch-up run for the most recent miss in the last seven days (Claude Code docs). Cloud sessions keep going even if you close the app or shut the computer down (Claude Code docs).

Why the lock column says what it does: Anthropic's Cowork and desktop pages don't mention locking. For Dispatch and computer use they say your computer needs to be awake with the Claude Desktop app open, and a locked computer is still awake. The Claude Code Remote Control docs go a step further and suggest switching phone notifications on when your screen locks, which only makes sense if Claude Code is still working behind the lock (Claude Code docs). What Anthropic doesn't say is whether computer use, browser use or Dispatch can do their clicking and typing while the screen is locked. If a task depends on those, test it before you walk away.

And a plain chat? Anthropic's new combined Claude experience, rolling out gradually to Pro and Max plans first, keeps more involved tasks running in the cloud if you close your laptop or leave the page (Claude Help Center). For an ordinary chat reply that's halfway written when the laptop sleeps, Anthropic's docs don't say.

Mac or Windows? Anthropic's rules are the same on both. What differs is the lid. On Windows, closing the lid is a setting. In Control Panel, under Power Options, the "When I close the lid" list sets what happens, and Windows supports Do nothing, Sleep, Hibernate and Shut down there, though Hibernate isn't offered on every PC (Microsoft Support, Microsoft Learn). Pick Do nothing and the laptop stays awake with the lid shut. On a MacBook, Apple documents lid-closed use with an external display and accessories connected, covered further down.

The Claude app also has a Keep computer awake switch, in Settings, then This computer, then System. It stops idle sleep. Anthropic's own docs spell out the limit: closing the laptop lid still puts it to sleep (Claude Code docs).

You set the idle timers yourself. On a Mac, the inactivity timers are in System Settings, under Lock Screen. On a Mac laptop, Battery, then Options, has a switch that stops automatic sleeping on the power adapter while the display is off (Apple Support).

One thing Hivra can't help with here: it doesn't run the Claude app or Cowork. What it runs is Claude Code, the terminal version, on a computer that stays on while you're on a paid plan. That's Fix 5 below.

Can I just keep the laptop awake?

For a short task at your desk, yes. Each command below holds the laptop awake only while Claude Code is running.

On Windows, set what closing the lid does in Control Panel, under Power Options, to "Do nothing".

caffeinate doesn't keep a MacBook awake with the lid closed. Apple documents closing the lid with an external display connected (Apple Support), and its dual-display guide lists an external keyboard and mouse or trackpad and power among the requirements (Apple Support). A laptop left running inside a bag can also get hot, and a reboot, a dropped network or a move to a different Wi-Fi network still ends the run. Keeping a laptop awake is a patch. If the run matters, put it on a machine that's meant to stay on. If you do go the laptop route, the keep-awake command builder writes the command for your setup and tells you what it won't cover.

Can I control it from my phone while the laptop is closed?

Depends which phone option you mean. Remote Control connects your phone or browser to a session that's still running on your computer, so that computer has to stay on and awake. It helps when you walk away from your desk. Shut the lid and leave, and the session sleeps with the laptop. The Remote Control guide has the commands, the two offline messages you can run into and how to keep the host machine online.

Claude Code on the web runs on Anthropic's infrastructure, so the session keeps going after you close your laptop and you can check on it from any device (Claude Code docs). A machine that stays on works too, reached over SSH or a browser view. On Hivra you can also send Claude Code work from your phone through the Telegram tab. Those runs execute on your Hivra computer.

Fix 1: tmux (or screen) on a machine that stays awake

Got an always-on machine (a desktop, a home server, a VPS)? Then tmux is the classic answer. It keeps your terminal session alive on that machine even after you disconnect. Anthropic's Remote Control docs recommend the same for a session on a remote machine: start it inside tmux or screen.

The tmux cheat sheet has the keys and commands on one page. If the CLI process itself exited and you need the chat history back, resume a Codex or Claude Code session covers the resume commands.

Your SSH connection can drop a hundred times and the session keeps working on the server. Reattach and you're back where the agent left off. The rule from the top still holds, though: tmux on your laptop doesn't survive sleep. It covers disconnects and closed terminal windows, and it can't do anything about a sleeping machine.

Prefer screen? Same idea, older tool:

tmux is the better default today. Panes, better scripting, active development. But screen ships preinstalled on more distros, and for this one job either works. To check your own setup for the usual ways these runs die, try the agent survival check.

Fix 2: nohup for one-shot headless runs

You don't always need the interactive session. Claude Code has a headless mode (claude -p) that takes a prompt, runs it and exits. For fire-and-forget jobs, nohup keeps that run alive after you log out:

nohup detaches the process from your terminal, so the logout SIGHUP never reaches it. That suits single tasks. It's the wrong tool for an ongoing interactive session, because you can't reattach and steer. Use tmux for that.

Fix 3: the DIY VPS route

No always-on machine at home? Rent one. Anthropic's system requirements for Claude Code list 4 GB of RAM, and a 4 GB server from a budget host runs roughly $5-10/month. Setup looks like this:

Prefer npm? npm install -g @anthropic-ai/claude-code installs the same binary and needs Node.js 22 or later. Then detach, close your laptop and go to sleep. The agent keeps working, and you can reattach from your phone over SSH if you want to check in.

Beyond the server rent, the DIY route has other costs. Setup takes an hour or two if the steps go clean, and more if they don't. The login flow on a headless server means copying URLs and codes between machines. Then comes upkeep: OS updates, CLI updates, a disk filling up with logs. Small jobs, but they land on you at bad times.

Security is yours too. The server sits on the public internet with your logged-in agent on it. SSH keys only, firewall on, fail2ban is a good idea. That's your job now. All you get is a terminal over SSH: no browser view of what the agent is doing, no file browser, nothing on mobile beyond a raw shell.

Fix 4: Anthropic's own cloud (Claude Code on the web)

If you only use Claude Code and your work lives in GitHub repositories, this may be all you need. A cloud session runs on machines Anthropic manages, and Anthropic's documentation says it keeps running after you close your laptop (Claude Code docs). You start one from claude.ai/code, from the Claude app, or from the terminal with claude --cloud.

The same docs set a few conditions. Cloud sessions are available on Pro, Max and Team plans, and to Enterprise users with eligible seats. The session clones a GitHub repository. You can send a local repository up as a bundle, but results can only be pushed back to GitHub.

Idle sessions expire. A session stops after a period of inactivity and the machine is reclaimed. You can reopen it with its history, but work that was still running in the background isn't restored. Your own files, tools, logins and other agents stay behind too, since it's Anthropic's environment.

A computer of your own makes sense when you want those things to stay put between tasks, or to run other agents beside Claude Code.

If you use Codex too, you've got the same choice with OpenAI's cloud. Codex cloud or your own computer lines the routes up with dates, and says when OpenAI's cloud is enough. Using Cursor? Same question, different answer. Cursor Cloud Agents covers what it costs to let Cursor work with the lid shut, and Cursor vs Claude Code helps you pick between the two.

Fix 5: a computer of your own on Hivra

Hivra runs the official Claude Code CLI from Anthropic on a private virtual machine set up for you, a computer whose only job is running your agent. Hivra's chat runs the CLI with permission prompts bypassed by default inside that VM, and the Permissions setting on the agent's Manage tab narrows that to Limited or Read-only. The skip-permissions guide lays out what bypassed prompts mean for your files, and which checks still run.

After launch, you sign in with your own Anthropic account on the computer, the same login flow as on your laptop. The login is stored on that VM. Hivra sets up the machine and runs it for you, so there's no server admin on your side.

In practice:

• On a paid Hivra plan the computer stays on and keeps your files, sessions and login. A Claude Code or Codex run you start inside tmux in the computer's Terminal tab keeps going after you close the laptop. On Claude Code, work you send through Telegram, connected in the agent's Telegram tab, runs on the computer, not in your browser.
• Connect your own bot in the agent's Telegram tab, under Manage, and send Claude Code work from your phone.
• You reach the computer from a browser. Chat, terminal, files, skills and browser tabs wrap the standard CLI, so you can check on it from your phone.
• There's a live self-hosted browser too. The Claude Code computer ships browser automation, so the agent can drive a real Chrome browser on its own VM.

Launch it from the Claude Code agent page.

Paid plans start at $9.99 a month for 2 vCPU and 4 GB of RAM, stay on without being paused for inactivity, and come with a 7-day money-back guarantee on card payments. The $9.99 plan is 2 vCPU and 4 GB of RAM, enough for Claude Code with its browser, and the pricing page has the full details. Before you commit, check whether your current Claude subscription covers round-the-clock usage. The Claude Code plan calculator does that math, and Claude Code pricing: Pro vs Max sets out which plan fits and what Anthropic publishes about the limits. Hivra is independent and is not affiliated with Anthropic.

DIY vs managed, side by side

If you like running servers and want total control, DIY is a fine route, and it's cheap. If you want the agent working tonight without owning another Linux server, managed gets you there faster.

Which fix should you pick?

• You use the Claude app, not the terminal: on Pro and Max plans, new Cowork tasks already run in Anthropic's cloud. Keep the app open on an awake computer only when a task needs your files, browser or apps.
• Short task, you're at your desk: caffeinate -i claude and leave the lid open.
• You already have an always-on machine: tmux. Done. Costs nothing.
• One long task has to finish overnight: nohup claude -p on any machine that stays awake.
• You only use Claude Code on GitHub repositories: try Claude Code on the web first.
• You want a permanent 24/7 setup and you like sysadmin work: a small VPS plus tmux, roughly $5-10/month.
• You want a permanent 24/7 setup without the sysadmin work: run Claude Code on Hivra. Sign in with your own Anthropic account and the computer stays up. Start long runs inside tmux in its Terminal tab, or send them through Telegram, so they keep going after you close the laptop.